AI detector Pangram flags two prize-winning stories, but both literary prizes doubled down

TuhinChakr · x · 2026-11-18

Blogger Yudhanjaya Wijeratne defends the AI-writing detector Pangram after controversies: in May 2026 it flagged the Commonwealth Prize-winning story as 100% AI-written, and in November 2026 the Atlantic used it to question a Caine Prize winner. Both prizes dismissed the findings despite obvious AI tells — the Caine Prize's own statement reads as AI-generated. The author argues critics lack technical understanding, that untrained humans are poor at spotting AI text, and that dismissing detection tools leaves writing prizes open to exploitation.

Original post →

More from AGI Musings

AGI Musings channel →